18.04 S18 Topic 12: Laplace Transform
Topic 12 Notes Jeremy Orloff
12 Laplace transform
12.1 Introduction
The Laplace transform takes a function of time and transforms it to a function of a complex variable �. Because the transform is invertible, no information is lost and it is reasonable to think of a function � (�) and its Laplace transform � (�) as two views of the same phenomenon. Each view has its uses and some features of the phenomenon are easier to understand in one view or the other. We can use the Laplace transform to transform a linear time invariant system from the time domain to the �-domain. This leads to the system function �(�) for the system –this is the same system function used in the Nyquist criterion for stability. One important feature of the Laplace transform is that it can transform analytic problems to algebraic problems. We will see examples of this for differential equations.
12.2 A brief introduction to linear time invariant systems
Let’s start by defining our terms. Signal. A signal is any function of time. System. A system is some machine or procedure that takes one signal as input does something with it and produces another signal as output. � � � � Linear system. A linear system is one that acts linearly on inputs. That is, 1( ) and 2( ) are inputs � � � � � � to the system with outputs 1( ) and 2( ) respectively, then the input 1 + 2 produces the output � � � �� �� 1 + 2 and, for any constant , the input 1 produces output 1. � � � � � � � � This is often phrased in one sentence as input 1 1 + 2 2 produces output 1 1 + 2 2, i.e. linear combinations of inputs produces a linear combination of the corresponding outputs. Time invariance. Suppose a system takes input signal � (�) and produces output signal �(�). The system is called time invariant if the input signal �(�) = � (� − �) produces output signal �(� − �). LTI. We will call a linear time invariant system an LTI system. Example 12.1. Consider the constant coefficient differential equation
3� ′′ + 8� ′ + 7� = � (�)
This equation models a damped harmonic oscillator, say a mass on a spring with a damper, where � (�) is the force on the mass and �(�) is its displacement from equilibrium. If we consider � to be the input and � the output, then this is a linear time invariant (LTI) system. Example 12.2. There are many variations on this theme. For example, we might have the LTI system
3� ′′ + 8� ′ + 7� = � ′(�), where we call � (�) the input signal and �(�) the output signal.
1 12 LAPLACE TRANSFORM 2
12.3 Laplace transform
Definition. The Laplace transform of a function � (�) is defined by the integral ∞ � � −��� � ��, ( ; ) = ∫ e ( ) 0 for those � where the integral converges. Here � is allowed to take complex values. Important note. The Laplace transform is only concerned with � (�) for � ≥ 0. Generally, speaking we can require � (�) = 0 for � < 0. Standard notation. Where the notation is clear, we will use an upper case letter to indicate the Laplace transform, e.g, (� ; �) = � (�). The Laplace transform we defined is sometimes called the one-sided Laplace transform. There is a two-sided version where the integral goes from −∞ to ∞.
12.3.1 First examples
Let’s compute a few examples. We will also put these results in the Laplace transform table at the end of these notes. Example 12.3. Let � (�) = e�� . Compute � (�) = (� ; �) directly. Give the region in the complex �-plane where the integral converges.
∞ ∞ (�−�)� ∞ �� � �� −�� �� (�−�)� �� e (e ; ) = ∫ e e = ∫ e = � − � 0 {0 0 1 if Re(�) > Re(�) = �−� divergent otherwise
The last formula comes from plugging ∞ into the exponential. This is 0 if Re(� − �) < 0 and undefined otherwise. Example 12.4. Let � (�) = �. Compute � (�) = (� ; �) directly. Give the region in the complex �-plane where the integral converges.
∞ −�� ∞ �� �e (�; �) = �e− �� = ∫ � {0 − 0 � if Re(�) > 0 = � divergent otherwise
The last formula comes from plugging ∞ into the exponential. This is 0 if Re(−�) < 0 and undefined otherwise. Example 12.5. Let � (�) = �. Compute � (�) = (� ; �) directly. Give the region in the complex �-plane where the integral converges.
∞ −�� −�� ∞ �� �e e (�; �) = �e− �� = − ∫ −� �2 {0 0 1 if Re(�) > 0 = �2 divergent otherwise 12 LAPLACE TRANSFORM 3
Example 12.6. Compute (cos(��)).
Solution: We use the formula e��� + e−��� cos(��) = . 2 So, 1∕(� − ��) + 1∕(� + ��) � (cos(��); �) = = . 2 �2 + �2
12.3.2 Connection to Fourier transform
The Laplace and Fourier transforms are intimately connected. In fact, the Laplace transform is often called the Fourier-Laplace transform. To see the connection we’ll start with the Fourier transform of a function � (�). ∞ �̂ � � � −��� ��. ( ) = ∫ ( )e −∞ If we assume � (�) = 0 for � < 0, this becomes
∞ �̂ � � � −��� ��. ( ) = ∫ ( )e (1) 0 Now if � = �� then the Laplace transform is
∞ � � � �� � � −��� �� ( ; ) = ( ; ) = ∫ ( )e (2) 0
Comparing these two equations we see that �̂(�) = (� ; ��). We see the transforms are basically the same things using different notation –at least for functions that are 0 for � < 0.
12.4 Exponential type
The Laplace transform is defined when the integral for it converges. Functions of exponential type are a class of functions for which the integral converges for all � with Re(�) large enough. Definition. We say that � (�) has exponential type � if there exists an � such that � (�) < �e�� for all � ≥ 0. Note. As we’ve defined it, the exponential type of a function is not unique. For example, a function of exponential type 2 is clearly also of exponential type 3. It’s nice, but not always necessary, to find the smallest exponential type for a function. Theorem. If � has exponential type � then (� ) converges absolutely for Re(�) > �. Proof. We prove absolute convergence by bounding